Hello,
I encounter an interesting problem while running the executable using the latest version of lammps (lammps-18Feb).
I am compiling lammps in linux (suse 11.2) platform.
The problem is the following.
I tried to do some print statements in the meam library file (meam_setup_done.F) and compiled using ifort (make -f Makefile.ifort)
After that, including the library, i compiled the lammps executable (lmp_linux)
While running the code, I do get the print statements, however, all the quantities printed using thermo_style are showing ‘nan’
except time step.
If I remove the print statement (write(1,*) ‘j,nr’, j, nr, in the loop over r values in compute_pair_meam subroutine) the code compiles and run properly.
I also like to mention that it doesn’t matter where I do the print statement, the same error occurs. Also, the Pair time (%), etc will be different without the print statement.
Could you suggest me what to do.
Thanks,
Maya
Hello,
I encounter an interesting problem while running the executable using the
latest version of lammps (lammps-18Feb).
I am compiling lammps in linux (suse 11.2) platform.
The problem is the following.
I tried to do some print statements in the meam library file
(meam_setup_done.F) and compiled using ifort (make -f Makefile.ifort)
After that, including the library, i compiled the lammps executable
(lmp_linux)
While running the code, I do get the print statements, however, all the
quantities printed using thermo_style are showing 'nan'
except time step.
If I remove the print statement (write(1,*) 'j,nr', j, nr, in the loop over
r values in compute_pair_meam subroutine) the code compiles and run
properly.
I also like to mention that it doesn't matter where I do the print
statement, the same error occurs. Also, the Pair time (%), etc will be
different without the print statement.
Could you suggest me what to do.
first of all, i would see, if this behavior is specific to your version
of the intel compiler, i.e. try different compilers or versions.
second, try to narrow down what exactly would generate the
unexpected behavior. what you are seeing is a symptom not
the real cause. particularly, you should start from clean sources
and then find the smallest modification that can generate
the unexpected behavior. lastly, fire up a debugger and
find out what is causing this for real. 
axel.
I would also suggest running a small problem under
a memory checker (e.g. valgrind) to see if it turns
up something.
Steve
Hi,
I think the problem is in the meam_setup_done.F file.
Any change in the compute pair meam subroutine cause this problem
If u give a print statement outside the loops, there is no issue.
Any clue?
why do you want me to blindly guess, when
i already gave you a list of steps to do, that
will help to narrow down the issue systematically.
for all i know, you may have just messed up
the file by accident and are causing memory
corruption this way.
axel.
p.s.: always send a copy of your replies to the list.